Diagnosing failure patterns in Deep Research Agents (DRAs) remains a critical challenge. Existing benchmarks predominantly rely on end-to-end evaluation, obscuring intermediate hallucinations that accumulate throughout the research trajectory. To bridge this gap, we propose a shift from outcome-based to process-aware evaluation by auditing hallucinations in the full plan-search-summarize trajectory. We introduce the PING Taxonomy, which categorizes DRA hallucinations into four complementary types: Propagation, Intent, Noise-induced, and Grounding. We further instantiate this taxonomy into a fine-grained evaluation framework that decomposes trajectories into atomic actions, claims, and sub-queries for rigorous verification, and we validate its reliability on standard fact-checking benchmarks and human-reviewed trajectories. Leveraging this framework to isolate 100 hallucination-prone tasks, including adversarial scenarios, we curate DeepHalluBench. Experiments on six representative DRAs show that, on our hallucination-prone stress-test set, all evaluated systems still exhibit non-negligible reliability gaps. Furthermore, our diagnostic analysis traces these failures to systemic deficits, especially hallucination propagation and cognitive biases, providing actionable insights for future architectural optimization. Code and data are available at https://github.com/yuhao-zhan/DeepHalluBench.
Before the Token Commits: Trajectory-Level Benchmarking of Visual Hallucinations in Diffusion VLMs
Nanjing University of Aeronautics and Astronautics(南京航空航天大学);State Key Laboratory of Ocean Sensing, ZJU-Hangzhou Global Scientific and Technological Innovation Center, Zhejiang University(浙江大学杭州国际科创中心海洋感知全国重点实验室);Security Capability Center(360安全能力中心)
Why Fine-Tuning Encourages Hallucinations and How to Fix It
Hebrew University of Jerusalem(海法大学);University of Illinois Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校);Technion – Israel Institute of Technology(技术学院 – 以色列理工学院);University of Southern California(南加州大学)
AI Hallucination from Students' Perspective: A Thematic Analysis
Department of Computer and Information Engineering, Center for Cyber Physical Systems, Khalifa University(计算机与信息工程系,跨物理系统中心,哈利法大学);Department of Psychology, University Malaysia Sabah(心理学系,马来西亚Sabah大学)
CHEM: Estimating and Understanding Hallucinations in Deep Learning for Image Processing
Munich Center for Machine Learning (MCML)(慕尼黑机器学习中心);University of Tromsø(特罗姆斯大学);German Aerospace Center (DLR)(德国航天中心);Foundation for Research and Technology Hellas (FORTH)(希腊研究与技术基金会)
Beyond the Final Answer: Evaluating the Reasoning Trajectories of Tool-Augmented Agents
Graduate School of Data Science, KAIST, Daejeon, South Korea(数据科学研究生院,韩国科学技术院,大田,韩国);Department of Industrial and Systems Engineering, KAIST, Daejeon, South Korea(工业与系统工程系,韩国科学技术院,大田,韩国);Department of Artificial Intelligence, Yonsei University, Seoul, South Korea(人工智能系,延世大学,首尔,韩国)